These are attacks that strip a model's value. By probing how the model reasons (adversarial reasoning extraction, which exposes inner chain-of-thought), an attacker gains what's needed to distil an unauthorized copy of the model, leaking its proprietary reasoning and intellectual property.
Glossary topic
Model reasoning and IP extraction
Terms in this topic
Track this in the live feed
See how this plays out in real AI security and governance developments.
Open the feed →